- PE: Memory leak
- PE: Do not initiate fencing for unclean nodes when fencing is disabled
- crm_node: Only remove the named resource from the cib
- pengine: Support site-specific instance attributes based on rules (FATE#316118)
- pengine: Ignore failure-timeout only if the failed operation has on-fail="block" (bnc#871175)
- Upstream version cs: f7859e2e06c8cf320482c3e4b447ddbe0a96fe83
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=129
* bug-812269_pacemaker-fencing-device-register-messages.patch
- Build: Update versions in configure.ac and sample spec file
- fencing: Do not unregister/register all stonith devices on every resource agent change
- PE: Unfencing is based on device probes, there is no need to unfence when normal resources are found active
- pengine: handles edge case where container order constraints are not honored during migration
- pengine: Allow container nodes to migrate with connection resource
- xml: Resolve memory leak when logging xml blobs
- PE: Do not unfence nodes that are offline, unclean or shutting down
- PE: Ensure unfencing only happens once, even if the transition is interrupted
- crmd: Actions can only be executed if their pre-requisits completed successfully
- crmd: Do not erase the status section for unfenced nodes
- Fencing: Correctly record which peer performed the fencing operation
- lrmd: Expose logging variables expected by OCF agents
- logging: Simplify processing of logging directives
- logging: daemons always get a log file, unless explicitly set to configured 'none'
- PE: Ensure unfencing occurs before fencing devices are (re-)probed
- fencing: Filter self-fencing at the peers to allow unfencing to work correctly
- fencing: Automatically switch from 'list' to 'status' to 'static-list' if those actions are not advertised in the metadata
- lrmd: Provide stderr output from agents if available, otherwise fall back to stdout
- fencing: properly detect new fencing devices for v2 cib updates
- acl: Prevent use-of-NULL
- cts: automatically determine whether to run RemoteBaremetal (bnc#856353)
- pengine: Correctly search failcount
- Upstream version cs: ff74a591e670dbdb86f7cddeec989167de521216
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=127
- xml: Correctly determine the lowest allowed schema version
- xml: Construct the table of supported schemas
- cib: All schema upgrade requests must go via the DC
- cib: Determine the best value of validate-with if one is not supplied
- acl: Ensure known users are not able to read the cib by default
- xml: Add missing schema file
- Upstream version cs: d9fbba56804a7a7bcceaa5d4b7e29e3f15fc0e59
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=124
- PE: Delay unfencing until after we know the state of all resources that require unfencing
- PE: Fencing devices default to only requiring quorum in order to start
- PE: Automatically re-unfence a node if the fencing device definition changes
- pengine: Only report 'migrate' in transition summary if migration is actually taking place
- crm_node: Set correct subsystem in node rm cache msg
- crm_node: Prevent use-after-free in tools_remove_node_cache()
- crmd: make resource ID case sensitive
- fencing: Correction of the registration from a difference.
- crm_report: Add support for ISO8601 time format in logs (bnc#870886)
- PE: Potential memory leak
- xml: Support resources that require unfencing
- crmd: make node_state erase correctly
- xml: Add the ability to have lightweight schema revisions
- xml: Split up the RelaxNG schema files so they're easier to version individually
- crm_get_msec: prevent integer overflow
- corosync: Reduce log severity for a redundant message (bnc#870624)
- attrd: Reduce log severity for normal messages (bnc#870624)
- cib: When data was old, it fixed so that the newest cib might not be acquired.
- crm_attribute: Default --lifetime to forever when using --node (bnc#870696)
- Upstream version cs: 76579e813b5aaa1972d400055233519f58bbc0c6
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=123
- crm_report: extract additional DLM lockspace data
- xml: If validate-with is missing, we find the most recent schema that accepts it and go from there
- tools: fix "crm_shadow --reset <name>"
- controld: Remove '-q 0' from default dlm_controld arguments
- acl: Authenticate pacemaker-remote requests with the node name as the client
- controld: Verify killall is available on resource start
- Prevent use-of-NULL spotted by coverity
- crmd: Correctly detect changes to the configuration root
- crmd: Fenced nodes that return prior to an election do not need to have their status section reset
- xml: Do not log deleted attributes
- xml: Correctly mark the parent's of deleted attributes as dirty
- crm_diff: Allow the generation of xml patchsets without digests
- crmd: Prevent memory leak when accepting a new DC
- xml: Find the minimal set of changes when part of a document is replaced
- xml: Create the correct diff when only part of a document is changed
- Upstream version cs: 61e95e4864be0301be528874eed65266f4e62cf0
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=120
- services_action_cancel: Interpret return code from mainloop_child_kill() correctly
- fencing: make hashtable mapping hostname->alias case insensitive for hostname
- attrd: Make hostname hashtable case insensitive
- membership: make caching peers by uname case insensitive
- crmd: make lrm_state hash table not case sensitive
- lrmd: clean up the agent's entire process group
- xml: Enable resource sets in location constraints for 1.2 schema (bnc#867880)
- Upstream version cs: 8bee850bb95ee01dcba03038ddfd65d22e195cf6
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=119
- lrmd: Cancel recurring operations before stop action is executed
- lrmd: Merge duplicate recurring monitor operations
- services: Do not allow duplicate recurring op entries
- crmd: Ack pending operations that were cancelled due to rsc deletion
- crm_mon: Add failures to --as-xml output
- cib: Check if on-disk cib is corrupt in the separate on_disk_cib_corrupt() rather than in validate_on_disk_cib() (bnc#866618)
- crmd: Do not start timers for already completed operations
- acl: Read access must be explicitly granted
- xml: Better move detection logic for xml nodes
- cib: Do not disable cib disk writes if on-disk cib is corrupt (bnc#866618)
- xml: Correctly calculate the adjusted offset of xml elements
- crm_diff: Mirror how the cib operates for better testing
- cib: Ingore patch failures if we already have their contents
- cib: Prevent use-after-free when cleaning up after query operations
- xml: Mark the parents of a partial replace as dirty
- xml: Correctly enforce ACLs after a replace operation
- cib: Do not try to use an empty value for CIB_shadow
- crm_resource: Gracefully handle -EACCESS when querying the cib
- xml: Correctly observe ACLs when creating filtered copies of xml documents
- xml: Only test write permissions for attributes that are being changed
- xml: Always allow new scaffolding - node with no attributes or only an id field
- cib: Memory leaks caused by ACLs returning filtered copies
- xml: Set flags to indicate when updates were rejected by an ACL
- cib: Respect the CIB_user environment variable for ACLs
- xml: Unpack and apply ACLs when tracking changes
- cib: Implement CIB_file-mode ACLs for testing
- remote: Enable migration support for baremetal connection resources by default
- cibadmin: Allow priviliged clients to perform tasks as unpriviliged users
- cibadmin: Remove dangerous commands that exposed unnecessary implementation internal details
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=115
- systemd: Prevent use-of-NULL when determining if an agent exists
- xml: Fix segfault in TRACKING_CHANGES()
- PE: Memory leak when updating cancel operations
- xml: Prevent memory leak in xml _private field
- pengine: Performance increase, stack allocated score2char
- crm_report: Only gather dlm data if dlm_controld is running
- Fix: memory leak found by valgrind was corrected.
- Fencing: Add the ability to call stonith_api_time() from stonith_admin
- Fencing: Pass the correct options when looking up the history by node name
- cib: Ensure 'cibadmin -R/--replace' commands get replies
- xml: Correctly initialize and find v1-style patch version numbers
- crm_mon: Add -R switch to show clone details
- Build: Install compatibility.h
- fencing: Teach stonith-ng about finding device and topology changes in the v2 patch format
- crm_report: Additional dlm detail
- crm_report: Allow ssh alternatives to be used
- crmd: Correctly recognise operation cancellations we initiated
- crmd: Improved logging of transition cancellations
- crmd: Correctly detect lrm refreshes
- crmd: Ignore 'move' operations to the cib, we only care about content changes
- PE: Ensure instance numbers are preserved for cloned templates
- Bug cl#5196 - PE: Check values after expanding templates
- crm_mon: Ensure resource attributes have been unpacked before displaying connectivity data
- Upstream version cs: ca8234daf2bbb2d0fd2fb47cdbb6667f892081bb
- pengine: Support id-ref in nvpair with optional "name" (FATE#316118)
* pacemaker-pengine-nvpair-id-ref.patch
* pacemaker-pengine-nvpair-id-ref-test.patch
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=111
- pengine: Support per-resource maintenance mode (bnc#863071)
- pengine: cl#5187 - Prevent resources in an anti-colocation from even temporarily running on a same node (bnc#856135, bnc#862319)
- controld: handling startup fencing within the controld agent, not the dlm
- Upstream version cs: 0b7d85a159cc92e00ebded65c9a33edce7ba1c22
- cib: Erasing the cib is an admin action, bump the admin_epoch instead
- mcp: Tell systemd not to respawn us if we exit with rc=100
- crm_mon: Allow the file created by --as-html to be world readable
- crmd: Do not overwrite existing node state when fencing completes
- PE: Default sequential to TRUE for resource sets for consistency with colocation sets
- Bug rhbz#1036631 - Stop members of cloned groups when dependancies are stopped
- PE: Ensure stop actions are viable when reporting what will happen to a resource
- xml: Consolidate patchset logging and apply functionality
- cib: Validate that everyone still sees the same configuration once all updates have completed
- cib: Validate clients are correctly applying v2-style diffs at most once a minute
- xml: Allow both patch formats to include and validate an xml digest
- election: If an error occurrs in crm_uptime(), compare as {0,0}
- xml: Correctly compare the current feature set when determining which patch format to use
- crmd: Support failcount processing with the new patch format
- crmd: Handle patchset directives that change the configuration top-level
- cib: Make patchset validation a run-time switch
- cib: Implement zero-copy status update
- cib: Enable the more efficient xml patchset format
- crmd: Add support for parsing resource operations in the v2 patch format
- xml: New function for calculating patchset version details
- cib: Auto-detect which patchset format to use
- xml: Allow the source of xml library errors to be traced
- cib: Send all r/w operations via the cluster connection and have all nodes process them
- xml: Use an optimized xpath search function for applying patch sets
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=108
* Drop pacemaker-Wno-cast-align.patch which has been merged upstream
- cib: Use the correctly typed enum value during initialization
- Build: Add dbus as an rpm dependancy
- Fix: Bug rhbz#1057697 - Use native DBus library for systemd async support to avoid problematic use of threads
- remote: Fail remote-node connection rsc monitor ops early on unexpected disconnect
- tools: Return error when attempting to delete an unspecified attribute
- remote: Allow baremetal remote-node connection resources to migrate
- lrmd: Send event alerting estabilished clients that a new client connection is created.
- Fix: Bug rhbz#1057697 - Use native DBus library for systemd support to avoid problematic use of threads
- xml: Fix segfault in find_entity()
- Upstream version cs: 6290953f2b035a92a8f093973b94c5cf38b79cd9
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=106
- pengine: cl#5186 - Avoid running rsc on two nodes when node is fenced during migration
- cluster: Fix segfault on removing a node (bnc#858745)
- services: Reset the scheduling policy and priority for lrmd's children without replying on SCHED_RESET_ON_FORK (bnc#858857)
- fencing: Update stonith device list only if stonith is enabled (bnc#857779)
- services: Correctly reset the nice value for lrmd's children (bnc#858857)
- pengine: Correctly perform partial migrations when node's uname is not equal to the id.
- fencing: Immediately fail remote fencing operation on peer timeout
- cl#5055: Improved migration support.
- stonith_admin: Ensure pointers passed to sscanf() are properly initialized
- Fix: Prevent potential use-of-NULL
- crmd: Prevent memory leak in error paths
- services: Fixes segfault associated with cancelling in-flight recurring operations.
- pengine: cl#5174 - Allow resource sets and templates for location constraints (FATE#315158)
- Upstream version cs: 31d76eedee4aecfa9b204bd3006c0ce2c7b569a4
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=95
- pacemakerd: Fixed a problem that only pacemakerd did not inherit the logfacility of corosync.conf.
- pengine: Fixes memory leak in regex pattern matching code for constraints.
- crm_attribute: Do not swallow hostname lookup failures
- crmd: Avoid deleting the 'shutdown' attribute
- Upstream version cs: ad10acb1c4d668befd7ea1d39312356223d86615
- cpg: Correctly free sent messages (bnc#853675, bnc#856018)
- cluster: Correctly construct the header for compressed messages (bnc#851951, bnc#856018)
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=92
- crmd: Improve throttling response on clusters of single core machines (bnc#838358)
- crmd: Correctly observe the configured batch-limit (bnc#838358)
- crmd: Prevent over-polling of throttle inputs (bnc#838358)
- crmd: Enable cluster-wide throttling when the cib heavily exceeds its target load (bnc#838358)
- crmd: Send the first throttle update (bnc#838358)
- crmd: A more reliable method of looking up migration options (bnc#838358)
- crmd: Allow non-CIB based throttling to be disabled (bnc#838358)
- crmd: Honor load thresholds lower than the cib maximum (bnc#838358)
- crmd: Tell the DC our job limit instead of the number of cores (bnc#838358)
- crmd: Track the CIB's CPU usage (bnc#838358)
- crmd: Fold the enforcement of migration-limit into the throttling code (bnc#838358)
- crmd: Don't start throttling until we're over the target threshold (bnc#838358)
- crmd: Do not send throttle updates to versions that don't support them (bnc#838358)
- crmd: Tweak the throttling threshold multipliers (bnc#838358)
- crmd: Simplify option processing for throttling options (bnc#838358)
- crmd: Simplify the logic for calculating the number of available cores (bnc#838358)
- crmd: Make the per-node action limit directly configurable in the CIB (bnc#838358)
- crmd: Slow down recovery on nodes with IO load (bnc#838358)
- crmd: Make the utilization limit configurable (bnc#838358)
- crmd: Use the load on our peers to know how many jobs to send them (bnc#838358)
- crmd: Notify peers of our relative load (bnc#838358)
- crmd: Track CPU usage on cluster nodes (bnc#838358)
- crm_resource: Avoid whitespace in transition keys when cleaning up and failing resources (bnc#851308)
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=85
- pengine: Disable container node probes due to constraint conflicts
- ipc: Have the ipc server enforce a minimum buffer size all clients must use.
- Bug cl#5181 - corosync: Ensure node IDs are written to the CIB as unsigned integers
- Upstream version cs: f57ad6c17624c0cbfc58c46f0bad7e28453506bb
Add a corresponding bug number:
- cluster: Correctly construct the header for compressed messages (bnc#851951)
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=84
* pengine: Do not check the allocation of resources during a colocation preview (bnc#849595)
- election: Correctly hook up the timeout callback
- crmd: Improve throttling response on clusters of single core machines
- fencing: Wait until all possible replies are recieved before continuing with unverified devices
- Bug cl#5179 - Corosync: Attempt to retrieve a peer's node name if it is not already known
- Core: Allow blackbox logging to be disabled with SIGUSR2
- PE: Correctly handle demotion of grouped masters that are partially demoted
- crmd: When the DC gracefully shuts down, record the new expected state into the cib
- crmd: When a peer expectedly shuts down, record the new join and expected states into the cib
- attrd: Memory leak
- attrd: Do something sane when older non-atomic versions send us messages
- crmd: Correctly observe the configured batch-limit
- attrd: Improve compatibility with older attrd versions
- controld: Use the correct variant of dlm_controld for corosync-2 clusters
- lrmd: Correctly calculate metadata for the 'service' class
- crmd: Prevent over-polling of throttle inputs
- crmd: Enable cluster-wide throttling when the cib heavily exceeds its target load
- Upstream version cs: 5a69be5ff56ec7fa66bcf9c2069af6211e4ebc98
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=83
* bug-806256_pacemaker-log-level-notice.patch
- Drop the obsolete patch:
* bug-800323_pacemaker-crm_mon-save-cib.patch
- Probe for baremetal remote-node connection resources.
- remote: Add support for ipv6 into pacemaker_remote daemon
- election/fencing: Fix the casts that would result in incorrect values for big-endian systems (bnc#842915)
- Fencing: Apply correct score to the resource of group
- crm_resource: Wait for the correct number of replies when cleaning up resources (bnc#842204)
- cpg: Only process one message at a time to allow other priority jobs to be performed
- remote: Handle endian changes between client and server and improve forward compatibility
- crmd: A more reliable method of looking up migration options
- crmd: Allow non-CIB based throttling to be disabled
- crmd: Honor load thresholds lower than the cib maximum
- crmd: Tell the DC our job limit instead of the number of cores
- crmd: Track the CIB's CPU usage
- crmd: Fold the enforcement of migration-limit into the throttling code
- crmd: Don't start throttling until we're over the target threshold
- crmd: Do not send throttle updates to versions that don't support them
- ipc: Use the higher of the configured buffer size or the default
- crmd: Tweak the throttling threshold multipliers
- crmd: Simplify option processing for throttling options
- crmd: Simplify the logic for calculating the number of available cores
- crmd: Make the per-node action limit directly configurable in the CIB
- crmd: Tweak the throttling code based on migration-limit and/or LRMD_MAX_CHILDREN (bnc#838358)
- crmd: Slow down recovery on nodes with IO load
- crmd: Make the utilization limit configurable
- crmd: Use the load on our peers to know how many jobs to send them
- crmd: Notify peers of our relative load
- remote: Properly version the remote connection protocol
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=74
- crm_master: Avoid sending redundant "--node" options to crm_attribute (bnc#846360)
- ipc: Client buffer size must never fall below our max read size
- ipc: Ensure buffer environment variable is read before setting max buffer len
- fencing: Deep copy current topology level list on remote op
- Upstream version cs: afdce1faecf82c778ed81eb5eb9b7518d36a1160
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=69
- crmd: Using PCMK_ipc_buffer variable for remote-node proxy
- plugin: Reset the scheduling policy of daemons to SCHED_OTHER with nice value -10 (bnc#843647)
- ipc: Compress msgs based on client's actual max send size
- pengine: Do not probe for connection resources
- Upstream version cs: 490e695539630cfc158c5967a99237659fac5f30
Add the bug numbers for those that have been fixed:
- crmd: Everyone who gets a fencing notification should mark the node as down (bnc#842915)
- crmd: Initiate node shutdown if another node claims to have successfully fenced us (bnc#842915)
- Fencing: Restore the ability to manually confirm that fencing completed (bnc#843699)
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=67
- fencing: Do not broadcast suicide if the on action is being executed
- remote: Option to enable/disable probes on container resource remote-nodes
- lrmd: Correctly cancel monitor actions for lsb/systemd/service resources on cleaning up (bnc#842471)
- services: Fix the executing of synchronous actions
- Upstream version cs: 728ec77851540f9ce64ee1b1269d967841571b25
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=65
- PE: Allow location constraints to take a regex pattern to match against resource IDs
- Bug rhbz#1011618 - Consistently use 'Slave' as the role for unpromoted master/slave resources
- cibadmin: Report errors during sign-off
- Fencing: Observe pcmk_host_list during automatic unfencing
- services: Prevent use-of-NULL when executing service actions
- crm_resource: Provide a meaningful error if --master is used for primitives and groups
- ipc: Raise the default buffer size to 128k (bnc#838358)
- Upstream version cs: ecd44e642db3b65ee83aceed0e93376894a58c47
- plugin: stonith-ng needs membership updates (bnc#841010)
- controld: allow the stonith-enabled check to be overridden
(bnc#842186, fate#315195)
- attrd: Avoid infinite write loop for unknown peers
- tools: Detect orphaned remote-nodes when setting attributes
- controld: don't start dlm if stonith-enabled is false (bnc#842186, fate#315195)
- pengine: Probe container nodes
- crmd: Tell attrd when remote-node is disconnected
- attrd: Give remote-nodes ability to set attributes with attrd
- Upstream version cs: c68919f46bbaa96ec4ff2c770f82588280277fe6
- crmd: Expose "expected-quorum-votes" option in the metadata only for plugin-based clusters
- decode_transition_key: Older systems do not support '%ms' in
sscanf
- cib: Guard against overly long user specified strings
- crm_shadow: Simplify prompt generation and impose a sane limit on
length
- fencing: Skip empty notifications from the cib
- election: Do not update state when ACKs arrive
- cib: Additional logging for invalid diff updates
OBS-URL: https://build.opensuse.org/package/show/network:ha-clustering:Factory/pacemaker?expand=0&rev=64
